AI Innovation Institute

Abstract: The landscape of machine learning evolves rapidly and the complexity of the networks and their architectures defies easy comprehension. AI is touted as the next scientific revolution by allowing the processing and pattern-finding in increasingly massive data sets. One potential end results could be AI enhanced measurement technologies, but what does that mean? This talk will give examples of how classical tools indicate the technical obstacles to this vision in terms of understanding training processes, model comparisons, and feature embeddings. While the results in this talk are largely empirical, they point to interesting directions for (infomation?) theoretical investigation.

Bio: Anand D. Sarwate is an Associate Professor in the Electrical and Computer Engineering Department at Rutgers, The State University of New Jersey. He received B.S. degrees in math and electrical engineering from MIT and a Ph.D. in electrical engineering from UC Berkeley. Prior to joining Rutgers he was a Research Assistant Professor at TTI-Chicago and a postdoc at the ITA Center at UC San Diego. His research interests include information theory, machine learning, signal processing, optimization, and privacy and security.
Location: Light Engineering 250

CSE 600 Seminar Series | Fall 2025

Speaker: Tuhin Chakrabarty

AI Seminar: Computational Pathology: Deep Learning, Classification and
Predicting the Future  - Joel Saltz

Abstract:  Pathologists have been looking at tissue through microscopes since the 1800s.  During each pathologist's career,  he or she views slides having  roughly 1,000,000,000,000 cells. Deep learning methods are rapidly being developed to assimilate the huge amount of information walked inside of tissue images and to use this information to predict outcomes and responses to treatments.

Stony Brook is a leader in this type of multi-disciplinary work. I will provide an overview of Stony Brook computational Pathology efforts and articulate how these have the potential to create biomedical advances as well as to drive development of new computer science. 


Bio: Dr. Joel Saltz is a leader in research on advanced information technologies for large scale data science and biomedical/scientific research. He has developed innovative pathology informatics methods, including: the first published whole slide virtual microscope system; pioneering pathology computer-aided diagnosis techniques; and methods for decomposing pathology images into features and linking those features to cancer omics, response to treatment and outcome. He has broken new ground in big data through development of the filter-stream based DataCutter system, the map-reduce style Active Data Repository and the inspector-executor runtime compiler framework. He has also been an active contributor in clinical informatics, having developed
predictive models for hospital readmissions, point of care laboratory testing quality assurance systems, decision support systems for electrophoresis interpretation and graphical user interfaces to support clinical data warehouse queries. Dr. Saltz has been a pioneer in establishing the field of biomedical informatics; he founded and built two highly successful departments of biomedical informatics, one at Ohio State University and one at Emory University. In 2013, he came to Stony Brook as Vice President for Clinical Informatics and Founding Department Chair of Biomedical Informatics - to create a living laboratory for biomedical informatics and to create a third unique biomedical informatics department dually housed in the School of Medicine and the College of Engineering. Dr. Saltz is trained both as a computer scientist and as a physician through the MSTP program at Duke University. He has deep experience in computer science, having served on the computer science faculties at Yale University and the University of Maryland. He completed his residency in clinical
pathology at Johns Hopkins University and he is a practicing, board-certified clinical pathologist. 

Abstract: Robot control has evolved from optimization-based controllers---precise but task-specific---through deep reinforcement learning's learned policies, to Vision-Language-Action (VLA) models that leverage pretrained vision-language backbones for language-conditioned manipulation across diverse tasks.
Despite their promise, VLAs exhibit a critical limitation: they function primarily as trajectory learners rather than skill learners. Recent evaluations reveal that VLAs often fail when faced with even minor variations in object initialization or environmental conditions, suggesting they memorize specific trajectories rather than acquiring generalizable manipulation skills. Attempts to address this through 3D spatial representations have shown limited success, indicating that the missing component may be more fundamental than geometric understanding alone.
This work argues that World Models (WMs)---internal representations that predict future states given actions---constitute the missing piece for robust VLA systems. We present one completed contribution and two ongoing investigations.
We developed a dual-layer world model for human-robot interaction that anticipates both physical scene evolution and latent human preferences for assistive tasks. Building on these foundations, we present ongoing work probing VLA internal representations to verify implicit world model existence, and propose a WM-VLA integration approach operating in the native visual domain through embedding prediction and image decoding.
Together, these contributions and investigations establish a foundation for WM-VLA systems, pointing toward robust, generalizable robot policies.
Speaker: Jason Qin
Location: NCS 220

Abstract: Recent advances in Spatial Transcriptomics (ST) pair histology images with spatially resolved gene expression profiles, enabling predictions of gene expression across different tissue locations based on image patches. This opens up new possibilities for enhancing whole slide image (WSI) prediction tasks with localized gene expression. However, existing methods do not fully leverage the interactions between different tissue locations, which are crucial for accurate joint prediction. To address this, we introduce MERGE (Multi-faceted hiErarchical gRaph for Gene Expressions), which combines a multi-faceted hierarchical graph construction strategy with graph neural networks (GNN) to improve gene expression predictions from WSIs. By clustering tissue image patches based on both spatial and morphological features, and incorporating intra- and inter-cluster edges, our approach fosters interactions between distant tissue locations during GNN learning. As an additional contribution, we evaluate different data smoothing techniques that are necessary to mitigate artifacts in ST data, often caused by technical imperfections. We advocate for adopting gene-aware smoothing methods that are more biologically justified. Experimental results on gene expression prediction show that our GNN method outperforms state-of-the-art techniques on multiple metrics such as mean squared error (MSE), mean absolute error (MAE), and pearson correlation coefficient (PCC). Qualitative analysis establishes the effectiveness of MERGE in capturing cancer marker genes, thus consolidating its utility in diagnostics. As an extension of this work, we use MERGE in a setting with an uncertainty calibration branch to perform robust gene expression smoothing. We show that using patch-wise uncertainty from an uncertainty calibration model and the gene expression predictions from MERGE to enrich the ground truth gene expression matrix, results in better alignment with pathologist annotations, thus establishing that the smoothing is biologically informed.

Speaker: Aniruddha Ganguly

Location: Virtual Zoom Meeting


https://stonybrook.zoom.us/j/5474847973?pwd=Sng0Q2h1c1d3cm9sbFBmYUczMHZNdz09
Meeting ID: 547 484 7973
Passcode: 206739

Professor Nanpeng Yu from UC Riverside present Machine Learning and Big Data Analytics in Power Distribution Systems.

Abstract: The electric utility industry is being swamped by petabytes of data coming from various sources such as smart meters, phasor measurement units, SCADA systems, geographical information systems and customer management systems. The primary and secondary value embedded in the complex and heterogeneous data sets from power distribution systems is immense. However, algorithms and applications for unlocking the potential of big data in power systems are at an early stage of development. This talk discusses the recent advancement of machine learning algorithms and big data analytics methods in power distribution systems. In particular, we will explain how to develop hybrid algorithms, which synergistically combine the merits of state-of-the-art machine learning algorithms and physical model-based methods. We will take a deep dive into the following applications: network topology identification, electricity theft detection, estimation of behind-the-meter solar generation and data-driven distribution system controls.

Bio: Dr. Nanpeng Yu received his B.S. in Electrical Engineering from Tsinghua University, Beijing, China, in 2006. Dr. Yu received his M.S. degrees in Electrical Engineering and Economics and Ph.D. degree from Iowa State University in 2010. Before joining University of California, Riverside, Dr. Yu was a senior power system planner and project manager at Southern California Edison from Jan, 2011 to July 2014.

Currently, he is an Associate Professor in the Department of Electrical and Computer Engineering at the University of California, Riverside, CA. Dr. Yu is the recipient of the Regents Faculty Fellowship and Regents Faculty Development award from University of California. He received multiple best paper awards from IEEE Power and Energy Society General Meeting, IEEE Power and Energy Society Grand International Conference and Exposition Asia and the Second International Conference on Green Communications, Computing and Technologies.

Dr. Yu is the director of Smart City Innovation Laboratory at UC Riverside. He currently serves as the vice chair of the distribution system operation and planning subcommittee of IEEE Power and Energy Society and the co-chair for IEEE Big Data Applications in Power Distribution Networks Task Force. Dr. Yu currently serves as the associate editor for IEEE Transactions on Smart Grid and International Transactions on Electrical Energy Systems.

The Program in Writing and Rhetoric
Invites you to
A Rhetorical/Deliberative Framework for AI Language Model Alignment
featuring
Prof Zoltan Majdik Professor
North Dakota State University
In this talk, Prof. Majdik proposes a framework for aligning LLMs with values grounded in the norms of rhetorical culture and deliberative democracy. Alongside long-standing AI alignment value targets like safety and transparency, this AI alignment framework assesses to what extent a language model exhibits human and humane values that foster communicative engagement, and it codifies approaches to tuning existing models to better align with such values.

Location: Humanities 1008

Abstract: Recent work in NLP uses debates between multiple LLMs to arrive at a more accurate conclusion. Earlier chain-of-thought prompting also shows improvements in accuracy when the model is asked to provide step-by-step reasoning in its response. Many publications since have developed strategies to improve the reasoning of model output with the goal of generating a more accurate result. However, even when asked to provide problem solving steps, the content of the reasoning provided by models is not well studied for all tasks and sometimes contains errors or conflicting statements even when the final result is correct. In fact, when evaluated across reasoning tasks, evidence shows that LLMs are not learning how to reason but are instead mimicking relevant solutions from their training sets.
By studying and evaluating the argumentation that LLMs provide, we can determine factors that may benefit or hinder the model's ability to give a complete, cohesive, and thorough answer. While there are signs that LLMs pattern match, finding where, when, and why this fails is valuable, as there may be ways to help the model imitate solutions that are more relevant to the task it is attempting to solve. Determining when pattern matching is not enough could show an area of improvement for future generations of LLMs. This research may separately aid in work on human-(AI)agent and inter-agent interaction. Specifically, frameworks could be used to determine when and why other models or humans are convinced by LLM-generated responses and which argument methods cause other models to change their response. Our current research in systematic versus heuristic cues shows that large language models sometimes present systematic or heuristic reasoning patterns based on prompting. Future research aims to explore other methods of classifying argumentation.

Speaker: Kiera Gross

Joining link: https://meet.google.com/xae-ywpv-udo